Epidemiology of morbidity and mortality in US and Canadian recreational scuba diving Buzzacott, P. and Schiller, D. and Crain, J. and Denoble, P. Abstract Objectives: This study investigates morbidity and mortality suffered by divers in the US and Canada. Study design: Prospectively recruited probability weighted sample for estimating the national burden of injury, and a weighted retrospective survey for estimating exposure. Methods: The National Electronic Surveillance System (NEISS) and Canadian Hospitals

Objectives: This study investigates morbidity and mortality suffered by divers in the US and Canada. Study design: Prospectively recruited probability weighted sample for estimating the national burden of injury, and a weighted retrospective survey for estimating exposure. Methods: The National Electronic Surveillance System (NEISS) and Canadian Hospitals Injury Reporting and Prevention Program (CHIRPP) were searched for scuba diving injuries. The Divers Alert Network (DAN) diving fatality database was searched for deaths and Sports and Fitness Industry Association (SFIA) estimates for diving were obtained from annual surveys. Results: In the US there were an estimated 1,394 emergency department (ED) presentations annually for scuba-related injuries. The majority (80%) were treated and/or released. There were an estimated 306 million dives made by US residents 2006-2015 and concurrently 563 recreational diving deaths, a fatality rate of 0.18 per 10 dives and 1.8 per 10 diver-years. There were 658 diving deaths in the US 2006-2015 and 13,943 ED presentations for scuba injuries, giving a ratio of 47 diving deaths in the US for every 1,000 ED presentations. There were 98 cases of scuba-related injuries identified in the CHIRPP data. The prevalence of scuba-related injuries for patients aged 3-17 years was 1.5 per 10 cases, and the prevalence of scuba related-injuries to patients 18-62 years was 16.5 per 10 cases. Discussion: In Canada and the US only one out of every 10,000 ED presentations is due to a scuba–related injury. That there are 47 deaths for every 1,000 ED presentations speaks to the relatively unforgiving environment in which scuba diving takes place. At 1.8 deaths per million recreational dives, mortality in scuba diving is nonetheless relatively low.


INTRODUCTION
Scuba diving is a popular recreational pursuit enjoyed by millions of divers in the US and Canada, 1 though it exposes divers to stresses that sometimes result in injuries. The pressure exerted upon a diver increases by one atmosphere for every 10 metres of seawater depth, effectively doubling when a diver descends from the surface to 10 m depth in the sea. Once a diver is submerged and then breathing, gas (usually air) is delivered at a pressure equivalent to the ambient pressure at whatever depth the diver is at. Therefore, the deeper a diver dives, the faster the air in the scuba tank is consumed and the more gas is taken-up by the diver's tissues. Immersion also contributes to the physiological stresses a diver experiences.
Drowning is the leading cause of death in diving, often after running out of gas, followed by cardiac causes and arterial gas embolisms. 2 Barotrauma from expanding gas during ascent, or from compression of the air spaces within the diver during descent, are the most common injuries reported to Divers Alert Network (DAN) each year through the 24-hours diving Emergency Helpline. 3 DAN is a not-for-profit diving safety organization that provides assistance to divers in need of medical help, monitors diving injuries, studies their causes and provides training in first aid and diving medicine.
Decompression sickness (DCS) and arterial gas embolisms (AGE) are commonly referred to collectively as Decompression Illness (DCI). Other hazards faced by scuba divers include hazardous marine life encounters, trauma, sprains and strains from heavy scuba equipment, boating injuries, falls while wearing scuba equipment and a myriad of other injuries. The risk of diving injuries for the population is not available because of a lack of data. The incidence data for diving injuries among recreational divers is incomplete because they are not mandatory reportable, unlike in military or scientific divers. There are two main rates of interest in diving; the number of injuries or deaths per 10 5 dives, and per 10 5 divers per year. 4

Morbidity rates
Dive conditions and dive populations vary across the world and, thus, injury rates may vary too. Over the 12 year period 1995-2007 the number of divers treated for DCI in Australia numbered 3,558. 5 Over the same period participation in sport surveys suggested this amounted to a DCI rate of 10.7 cases per 10 5 dives. 5 In Canada the number of air fills sold within a defined region was used to estimate that 146,291 dives had been made between 1999-2000, during which time there were 14 known injuries, amounting to a rate of 9.7 cases per 10 5 dives. 6 A similar, earlier study at a military base in Okinawa noted 94 cases of DCI between 1989-1995 and an estimated rate of 13.4 cases per 10 5 dives. 7 In the warmer Caribbean a study conducted in 1989-1990 noted 77,680 dives and 7 cases of decompression sickness (DCS), giving a rate of 9.0 cases per 10 5 dives. 8 A 1990 British study of 36,434 divers found 6.7 case of DCI per 10 5 dives. 9 In Japan a survey of 3,078 recreational divers between 1996-2001 found a reported incidence of DCS at 5.2 cases per 10 5 dives and a combined incidence of ear or sinus barotrauma, or DCS, of 49.3 injuries per 10 5 dives. 10 Numerous smaller studies have reported rates outside the above range, possibly due to wider confidence intervals associated with small sample sizes. In diving morbidity research to date, efforts have focussed on diving cohorts, for example customers of dive businesses in a geographically defined area, or members of DAN. To our knowledge no previous research has extracted scuba diving injury cases from much broader public health data.

Mortality rates
Using sports survey data, between 2002 and 2006 the fatality rate in Australia was estimated at 0.57 per 10 5 dives. 5 This was lower than found after counting the number of air fills sold, in British Columbia 1999-2000 at 2.05 per 10 5 dives, or in Okinawa 1989-1995 at 1.3 per 10 5 dives. 6,7 At a popular flooded quarry in the United Kingdom mortality over a five year period was observed to be 2.9 per 10 5 visiting divers. 11 Among members of the Divers Alert Network 2000-2006, the overall mortality rate was found to be 16.4 deaths per 10 5 insured person-years but mortality per 10 5 dives was not available due to the lack of exposure data. 12 More recently, DAN Japan reported a lower membership mortality rate over 2004-2012 of 6.9 deaths per 10 5 member years, while in Australia, similar to the result above, a rate of 0.5 deaths per 10 5 dives was based on annual surveys of Australian residents. 3,13

Preventive efforts
Since the development and widespread adoption of the internet divers have greater access to diving safety information than at any time previously. Concurrent with the growth in internet access, scuba diving equipment has continued to improve (for example modern diving computers are now worn by the majority of divers), diver training is more widely available today than ever before, and professional dive guides can be located as easily as clicking a mouse. It is, therefore, possible that diving is safer today than ever. What is not known are the current rates of morbidity and mortality suffered by divers in the US and Canada each year. This study aimed to determine scuba diving related morbidity and mortality based on publicly available ED in the US and Canada.

METHODS
In the US the Consumer Product Safety Commission (CPSC) maintain the National Electronic Surveillance System (NEISS), a national register of Emergency Department (ED) presentations at around 100 hospitals in US and US Territories. The data are sample probability weighted to reflect the ~5000 EDs in the wider US and US territories, deidentified, and each year the previous year's data are made publically available through the CPSC website. NEISS data have been used to describe a wide variety of injuries such as those resulting from paintball guns, 14 surfing, 15 and burns. 16 NEISS data for 2006-2015 were obtained from the CPSC website (www.cpsc.gov/en) and imported into Windows Notebook as tab-delineated text. Product code 1275 identifies injuries related to scuba in the NEISS dataset. 17 Cases involving product code 1275 were identified using SAS ver 9.4 (SAS, Cary, NC). A search for the word "scuba" in the case description field identified eight further cases, and manually checking the case descriptions for cases identified by product code 1275 found eight misclassified cases which were excluded. National estimates arising from NEISS data are considered by the Centers for Disease Control and Prevention as unstable and potentially unreliable when:  the estimate is less than 1,200  or the number of records used is less than 20, or  the coefficient of variation (CV) exceeds 30%. 18 The Canadian Hospitals Injury Reporting and Prevention Program (CHIRPP) supplied data concerning scuba diving injuries in children and adults presenting at emergency Departments at 11 paediatric and five general hospitals across Canada between 1990-2015. Since commencing data collection in April 1990, CHIRPP data have been used to describe various injuries including those from curling, 19 off-road vehicle use, 20 and martial arts. 21 A fuller description of CHIRPP data collection methods, including the names of participating hospitals, can be found here. 22 Records with the following criteria were assessed for inclusion in the study:

•
Records with Factor code "1167 Scuba Diving" (Direct Cause, Mechanism Factor, Contributing Factor, Breakdown Factor, and/or Factor codes 1-6) • Cases in which the patient's narrative contained a relevant text string: SCUBA, S.C.U., S C U, PLONGEE SOUS, BREATHING APPARATUS, or BREATHING APP Cases associated with the following activities were excluded from the study, via case-by-case manual cleaning:   Table 2. National estimates were not made because cell-counts were nearly all <20. Where race was identified in NEISS, 93% were classes as white whereas the US Census Bureau identify 77% of the population as white, 23 (though the racial distribution among recreational divers is uncertain). Affected body parts are presented in Table 3. Once again raw prevalence is presented because the majority of cell counts were lower than 20.  Table 4 presents the disposition of patients following diagnosis. The majority (80%) were treated and released or released without treatment. Less than 1% were dead on arrival or died in the ED. It is unknown how many later died as a result of their injuries.  a Some information for the years 2011-2015 is still being entered into the electronic CHIRPP system. *Injury sustained while diving, or while entering or exiting the water. **Injury was related to scuba diving or its equipment but did not occur while diving or exiting the water. Examples include sunburn, falling over scuba tank in the garage, injury on scuba boat during scuba-related outing, injured while misusing/playing with scuba equipment, etc.

Morbidity and mortality rates
Given there were 306,174,386 dives made by 30,444,000 US residents 2006-2015 (Table 1) and during that period there were 563 recreational diving deaths, then the fatality rate among US recreational divers is 0.18 per 10 5 dives (95% CI 0.16-0.21) and 1.8 per 10 5 diver-years (95% CI 1.49-2.21). There were 658 total diving deaths in the US 2006-2015 (Table 1) and 13,943 ED presentations in the US for scuba-related causes, giving a ratio of 47 deaths for every 1,000 ED presentations (95% CI . That this figure is far higher than the raw, unweighted rate of 3/378 (7.9/1000) presented in Table 4 suggests the majority of diving fatalities occur outside of the ED. Given the relative rarity of scuba-related injuries in the CHIRPP, the number of cases captured can be subject to random variation due to a small sample size. As such, these results must be interpreted with caution. Fatal injuries are also under-represented in the CHIRPP database because the emergency department data do not capture people who died before they could be taken to hospital or those who died after being admitted. Among adults the scubarelated presentation rate was 16.5 cases per 10 5 , more than ten times that among children presenting to the ED. Recent reports from the SFIA suggest only around 15% of all US divers are younger than 18 years old. 1 If this were taken into account then the presentation rates among adults may have been closer between the USA and Canada. In Canada and the US it appears likely that around only one out of every 10,000 ED presentations is due to a scubarelated injury, and the monthly distribution of these follow both the distribution of diving fatalities and non-fatal incidents reported to DAN. 3 Not all presentations at the ER are included in the NEISS data, for example chronic diseases are not. Therefore, the true prevalence of scuba injuries among ER presentations is likely even lower than one in 10,000.

DISCUSSION
DAN staff a 24/7 medical services call-centre offering free consult assistance to receiving ER staff unfamiliar with diving injuries.
That 75% of divers presenting to the ED are male is slightly higher than other studies report, for example 64% among DAN members 12  The types of injuries seen at the ED are typical of those reported to DAN's emergency call center, 3 with ears clearly the most commonly affected body part (Table 3). It may be reassuring that four out of five victims of a scuba-related injury will be discharged either following treatment at the ED or without any treatment at all ( The mortality rate of 0.18 per 10 5 dives is less than half that reported in Australia, but differing methodology may account for this discrepancy, rather than any real difference in risk. 13 The mortality rate of 1.8 per 10 5 divers per year lies between those found in British Columbia and Okinawa by counting air fills. This rate is comparable to many other outdoor recreations but direct comparisons are not possible due to variation in data sources and methodologies. It is substantially lower than that reported among DAN-insured members who were largely based in the US and Canada but DAN members are known to average 25 dives per year which is 2.5 times the exposure found by SFIA, and they were significantly older than the divers in this study and age is a known risk factor for death while recreational scuba diving. 12 The process of certification for scuba diving is relatively uniform worldwide, with only minor regional differences. Once certified there are usually no requirements for the maintenance of that certification. Unlike, for example, a driver's license, there is usually no requirement for regular re-testing of skills.
The limitations of this study include that the NEISS employ a weighted sampling frame geared to represent the population of US hospitals, but that scuba diving is a geographically concentrated sport confined to locations with access to popular dive sites and, therefore, the NEISS may not accurately reflect the true distribution of scuba-related injuries across the US.
The actual mortality rate may be higher than reported here because it is not certain that every diving fatality that occurred in the US between 2006-2015 is included in the DAN fatality database. Many diving injuries do not result in presentation to the ED and, therefore, the overall burden of injury attributable to recreational scuba diving is likely greater than that resulting in presentations to the ED or death. Also, many of the injuries listed in Table 2 had raw counts of less than 20 cases and, therefore, estimates of their annual incidence cannot be reliably made. As with calls to the DAN Emergency Helpline, barotrauma and DCS were the most common injuries seen at the ED. Lastly, it should also be remembered that not all the dives estimated to have been made by US residents were likely made within the US and also that many dives are made in the US by visiting divers, and country of residence is not noted in the NEISS data. Table 1: Estimated number of divers, dives, fatalities and ED presentations in the US 2006-2015 Table 2: Diagnoses by raw number (n=378)